PostHog

Good fit if…

• ✓Product and engineering teams shipping features and measuring in-app behavior
• ✓Startups wanting session replay, funnels, and A/B tests without stitching five vendors
• ✓Teams comfortable with event instrumentation (SDK or autocapture)
• ✓Organizations evaluating self-hosted analytics for data control

Consider other analytics tools if…

• −Blog-only sites that only need referrer and pageview reports
• −Teams with zero engineering bandwidth for event taxonomy and SDK maintenance
• −Marketing teams that only care about Google Ads attribution — GA4 integrates deeper there
• −Simple revenue-per-channel attribution for Stripe SaaS — DataFast is more direct

PostHog's core is product analytics: funnels, retention, paths, trends, and user-level profiles (when identified). Autocapture can track clicks and pageviews on web with minimal code; serious teams still define named events for trustworthy metrics.

Unlike GA4, PostHog is built for SaaS product questions — activation, feature adoption, churn signals — not just marketing landing pages. SQL insights and HogQL extend power users.

Event volume drives cost on cloud; disciplined instrumentation keeps bills predictable.

PostHog bundles feature flags, session replay, and experiments — tools that often require Hotjar + LaunchDarkly + Optimizely separately. For early-stage teams, consolidation reduces integration tax.

Session replay helps debug UX friction; flags enable gradual rollouts; experiments tie metrics to variants. Each capability adds script weight and privacy review surface — disclose replay to users where required.

Not every team needs all modules on day one. PostHog lets you enable capabilities incrementally.

PostHog Cloud is the fastest path — sign up, install snippet or SDK, start capturing. Self-hosted (open source) suits teams with infra teams and strict data residency.

Self-hosting PostHog is a real platform operation: ClickHouse, Kafka, upgrades, and backups. Budget engineering time, not just license savings.

EU and US cloud regions exist on paid plans; confirm current region options on vendor docs before committing.

PostHog Cloud includes a free tier (1M events/month and limited replay in our directory snapshot — verify current limits on posthog.com/pricing). Paid usage scales with events, replay recordings, and feature flag requests.

Self-hosted open source removes cloud event billing but adds hosting cost. For marketing-only sites, PostHog is usually overkill and overpriced vs Plausible.

Use our pricing calculator with events (not just pageviews) for realistic estimates.

Typical setup (web SaaS):

1. Create a PostHog project and install the JavaScript snippet or framework SDK.
2. Enable or disable autocapture based on noise tolerance.
3. Define key events (signup, upgrade, core feature use) with stable names.
4. Optionally enable session replay and feature flags for targeted users.
5. Build funnels and dashboards; set up experiments when traffic allows.

Expect half a day to several days depending on event taxonomy maturity.

Common pitfalls: anonymous vs identified user merge issues, autocapture noise, enabling replay without performance testing, and unbounded custom properties.

• −Overkill for simple marketing analytics — complexity and cost vs Plausible/Fathom.
• −Requires engineering ownership for clean event schemas and identity rules.
• −Session replay raises privacy and performance considerations.
• −Google Ads native integration is not GA4-class.
• −Script bundle heavier than lightweight analytics — benchmark pending on AnalyticsVS.